Hacker News new | past | comments | ask | show | jobs | submit login

"Discussion interviews" can suck because they're a lie. You're still being examined, and now you have to pretend that you aren't being examined in addition to performing well.

Some of my best interviewing experiences have been when as part of the interview I ended up having a discussion about something. But the interview didn't explicitly start with that format in mind.

Some of the worst interviewing interviewing experiences that I've had is when they say that it will be a discussion, and it is, up to the point when they spring an algorithm question out of the blue... it feels so scummy and fake. Ask me about the algorithm if you want, but mixing your question into 40 minutes of discussing other things and pretending that you aren't examining me is a farce.

The intention seems to be to make the experience more authentic and it often ends up having the exact opposite effect.

If your criteria for hiring boil down to "did I like talking to this person", you're probably not hiring well and you're allowing all kinds of biases to influence your decision. If your criteria are specific but you're hiding them behind the pretense of "discussion", you're doing everyone involved a disservice.




You’re being examined at an interview, but you’re also examining the person you may choose to work for. If you don’t need to take the job you’re in a much more powerful position and you can have an honest discussion to come to a mutually beneficial arrangement.


I’ve never had the benefit of not needing a job while doing interviews. Doing interviews while I’m at a job isn’t viable for me.


Discussion style interviews aren't about pretending they're not interviewing you though.


The title is "Don't do interviews, do discussions". That's repeated in the main text. The author seems to be concerned about the feeling of "I am being evaluated". I think that's counterproductive because it's false - being evaluatated is the whole point of the conversation and it's better if both sides were honest about it instead of lying to each other and playing games.


In context, I take it to mean "don't do the normal format of interviewing, use discussions". No amount of phrasing is going to convince you that you aren't sitting in an interview. I wouldn't want to convince you you're not in an interview anyway, that seems pretty dishonest. There's just way more value in having a (albeit can be fairly technical) discussion rather than typical call/response style of interview that consists pretty much solely of "did you memorize what I'm looking for". Particularly because I don't expect the person I'm interviewing to be a master of everything, having a discussion can lead to some common ground where we can go in deep on some of your actual previous experience.


"Interview" is ambiguous here. No matter what you do, it's always true that you're interviewing the candidate in the "assessing a candidate" sense, but you don't have to do this by interviewing them in the sense of "question and answer format you'd see a journalist use".


The tone of an evaluation can change the experience and isn’t boolean.


Which is another way of saying "life isn't boolean" honestly.


Ah! What you describe drives me bonkers!

I’ve had interviews which start very casually, “we just want to have a high-level conversation, no crazy whiteboard coding or anything”, then, absolutely randomly in the middle of said “high-level” discussion, I get

> “well we did prepare a question to test your quantitative reasoning skills; so assume we have a bug that is detected by an analyzer with a false-negative rate of 0.3% and the probability of a bug existing at all is 7% and <sea of numbers> what’s the probability that a given test will report blah blah blah”

Of course this is communicated purely verbally.

> “This looks like a typical Bayes rule problem. Is that what you want me to calculate?”

> “Maybe, maybe not. That could be an approach. We just want the final percentage. Feel free to pull out your phone and use the calculator.”

Then of course, on the spot, I’m fumbling around with whether you divide by P(A) or P(B) or … to find P(B|A), and mucking around on a 4-function calculator like a dweeb.

So much for the “high-level discussion.”


I interview quite a bit of people. I always do it in a laid back, conversational style. When I ask technical questions, somewhere towards the middle or end of the session, I do it because I just want to know that the person I talked with understands some fundamentals (it's never a tricky question; just something basic like creating some threads, etc).

I talk with them about their previous work, stuff they're proud of, their hobbies and other things. I've recruited teams that excelled compared to their peers, and were certainly more fun than others.


threading is your basic question at the end of an interview?


Yes, but really simple. It can get progressively worse though. Like: 1. Make a loop which counts and prints up to N 2. Now make it run in a thread 3. Now make another thread which only prints once this var hits modulo X == 0 etc


Actual fizzbuzz-level questions! Kudos for keeping the tradition alive.


Exactly, if you lead with some obscure or polarizing hobby thinking its a casual discussion its a big mistake

“Hiking, trips to the beach” those are the answers. Any behavioral interview training will say the same

Everyone is lying (or actually boring and unambitious)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: